Content uploaded by Rifan Kurnia
Author content
All content in this area was uploaded by Rifan Kurnia on Feb 16, 2023
Content may be subject to copyright.
1
2
AI-based Authentication System in IoT environments – A Systematic
Review
Rifan Kurnia
King’s College London, UK
rifan.kurnia@kcl.ac.uk
Abstract
Cybersecurity is essential for a solid foundation in the age of digital. The advancement of recent
technologies including Internet of Things (IoT) environments has brought the world to a new level of
interconnection. However innovative technologies always come with a new kind of vulnerability. An IoT
device can integrate with various devices through internet connections to provide intelligent services and
must protect its user privacy. It also has to provide security against various attack schemes in the
authentication system.
To improve the security of IoT authentication system, various research has been conducted in AI based IoT
authentication domain. The study tries to address the following research questions: 1) What are the common
challenges in an IoT environment, especially related to authentication problems? ; 2) How could AI help
IoT devices secure the authentication process? ; 3) How is the recent research trend in AI to solve the
authentication problems?.
To address the questions, systematic review of articles published in articles, conference proceedings, and
journals for the period of 2017–2022 is conducted.
Index Terms
AI-based IoT, IoT cyber-attacks, Cybersecurity, Anomaly detection, Authentication
I. Introduction
3
A. Background on IoT
Over the past two decades, the Internet has been constantly evolving. The beginning of the Internet
was marked by the World Wide Web (www) development, a network of linked web pages which are HTML
formatted on the internet. This network of static HTML pages has gradually evolved into Web 2.0, where
two-way communication has become ubiquitous, allowing users to participate, collaborate, and interact.
Whitmore et al. [4] explained that While Web 2.0 currently dominates the Internet, researchers say
they are working toward another objective, commonly cited as the Semantic Web or Web 3.0. The goal of
Web 3.0 is to make web content machine readable and marked up in a way that makes search engines and
machines work more intelligently. By tagging web content in a standardized format, machines can process
and share data without the need for human intermediaries. Along with the development of Internet
technology, the technology of short-range wireless communication using sensor networks and RFID tags
has also been developed. The fusion of his two technologies, the internet and sensor network, brings new
possibilities and perspectives. The framework’s ability to enable direct machine to machine communication
over the internet has prompted researchers to consider the benefits of connecting more machines to the
network, allowing them to connect to the internet as a giant network. This concept has spawned a paradigm
called the Internet of Things (IoT).
Kornaros et al. [3] discussed that IoT devices are mostly restricted devices, with poor tamper-proof,
for example, allowing modified firmware to access authentication data so that connected devices cannot
access personal data. may be leaked. With the rapid development of Internet of Things (IoT) devices and
cloud systems, the attack surface area of most cyber-physical systems is expanding rapidly, so there is little
that can be done to prevent the increasing volume and complexity of attacks.
As shown in Figure 1, IoT-enabled cyber-physical systems (CPS) for factories, smart grids, and
automobiles come with various communication interfaces and remote monitoring capabilities. Industrial
IoT networks have a large attack surface, making covert channels more difficult to detect. As a result, most
intrusion detection systems rely on the uniqueness of signatures or untampered signatures.
McNett et al. [7] showed in their journal that the number of internet-connected devices in the
industrial environment and their high-speed connections lead to countless network intrusion opportunities,
the threat is real. Altman Vilandrie and Company, a strategy consulting firm, reported in a 2017 survey that
46% of IoT security buyers have experienced an IoT-related security breach or breach in the past two years.
As nearly half of the organizations surveyed were affected by an IoT-related security breach, addressing
the issue became a top priority. The monetary impact of these offenses must also be understood. In the same
release, the firm researched that the average financial loss of an IoT security breach for a company valued
4
at less than $5 million was 13.4% of the company's revenue. The impact of an IoT security breach grows
with the size of the business affected.
Figure 1. An illustration of threat model in IoT infrastructure that involves a wide spectrum of cyber-attacks
[3].
Since their introduction in the late 1990s, IoT devices have been manufactured without much
attention to security. The indivisible problem of the absolute nature of end devices remains the biggest
bottleneck to achieving security in IoT. Security is a well-known challenge in IoT, especially in
authentication and authorization, and it remains one of the biggest security problems due to the limited
computation, storage, and control issues of end devices. secure and adaptive verification and approval
systems in. On the other hand, artificial cognition and machine learning are known to solve customization
and execution problems in the computing world. This white paper systematically reviews existing
documents that attempt to solve the verification security problem using AI and machine learning
approaches.
5
Figure 2. Authentication framework based on behavioral signals [11].
As shown in Figure 2, authentication framework in IoT can be categorized into four layers as follows.
1. Behavioral signals. These are a set of characteristic behavioral patterns or features that a decision-
making system can use to determine an individual's identity.
2. Unobtrusive sensing. This summarizes viable acquisition strategies for capturing available sensor
and behavioral cues.
3. Continuous computing highlights the workflow for authentication. This is to determine if the user
is authorized to access the IoT device (authentication) and to identify who the current user is.
4. Applications. These are typical scenarios where behavioral biometrics-based authentication is
applied.
B. Contribution of Paper
One of the exciting developments in technology is IoT. The usage trend of IoT devices is increasing
both in daily life and in industrial applications. While the development is massive, one of the biggest
concerns and challenges is how the IoT environment can secure user data.
This study is intended to identify what are the common cyber-attacks in IoT environments and will
focus on the authentication aspect. Other than that, the study aims to answer how AI can effectively help
IoT devices to secure the authentication process. For this purpose, readers will have an idea about IoT
6
security aspects and AI that can help quickly detect the attack. In addition, opportunities for future research
are also discussed to expand the perspective on the technological advancement in the area.
C. Paper Outline
This systematic review is structured as follows. In Section I the paper shares the introduction of IoT
to give the reader general insights on IoT development, importance, and trend. In Section II the paper
describes the methodology used in this systematic review, which consists of research questions, selection
methodology, and screening process. In Section III the paper discusses the analysis and results from the
systematic review. The readers will have a general idea of IoT security risks, challenges with previous
studies on IoT authentication, general AI process, and recent studies in AI-based IoT authentication. In
Section IV the paper will discuss the research questions based on the selected literatures. In Section V the
paper will discuss promising future research directions before concluding the review in Section VI.
II. Literature Review Methodology
A. Research Questions
The PICO(C) framework is used to help with the search strategy.
• Population: the type of Internet of Things
• Interventions: the AI methods to detect cyber-attacks in IoT
• Comparison: the difference if IoT infrastructure does not occupy with security detection aspects
• Outcomes: the opportunity to advance IoT with AI based cyber-attacks detection systems and the
potential future research in the area
• Context: the different research between academia on how to handle security challenges in IoT using
AI
Based on this method, the following research questions are constructed to further review collected
literatures.
1. What are the common challenges in an IoT environment, especially related to authentication
problems?
2. How could AI help IoT devices secure the authentication process?
3. What are the recent research trends in AI to solve authentication problems?
7
To address these research questions, a set of keywords are constructed to discover related
cybersecurity challenges in IoT environment and how AI could help to solve the challenges. The search
queries used are: IoT cyber-attacks, IoT security, IoT authentication, AI, machine learning.
B. Selection Methodology
The search methodology follows the steps proposed in:
1. Define the research questions. This is defined in the previous section based on PICO(C).
2. Break research questions into key phrases. The research questions can be broken down into
phrases such as IoT cyber-attacks, IoT security, IoT authentication, AI, machine learning., etc.
3. Identify the type of literature you require. The search focused on journals, conference
proceedings, books, and magazines as literature materials.
4. Define search criteria. The key phrases can be the search criteria to further sharpen the search.
5. Identify the most appropriate data sources. This research uses various literature databases, such
as IEEE, ACM, and Springer.
6. Combine the key phrases to automatically identify from Titles and Abstracts. Boolean
operator AND/OR is used to search the relevant literature.
7. Apply the search.
8. Analyse the result. From the search result, the truly relevant literature was analyzed and included
in the study, while excluding the others.
9. Document the search. Since systematic literature cannot be done in a night, a good documentation
of literature list based on the search is essential.
As well as the inclusion criteria, I also used the following exclusion criteria (EC) to help with the
search. The paper can be excluded if one of following criteria is satisfied.
• EC 1: the paper older than 2017.
• EC 2: the paper has no title nor abstract.
• EC 3: the paper is not presented in English.
• EC 4: the paper full text is not accessible.
8
C. Screening Process
This section is intended to elaborate the detailed process on the literature search and screening into
the refined list of literature that is truly relevant for the systematic review.
1. Papers identified from the KCL library search based on keywords (n = 153)
2. Papers identified from the manual screening (n = 57)
3. Papers identified that can be accessed in full text (n = 45)
The search was primarily performed in the KCL library database which taps into major literature databases
such as IEEE, ACM, and Springer.
Chart 1 depicts the research trend in the relevant topics. It can be concluded that the topic is getting
popular and is an interesting topic to explore further in the coming years.
Chart 1. Research trend of AI-based authentication on IoT
III. Analysis and Results
A. Cyber Attacks on IoT
For each layer in IoT, the need for security is very important and should consider the requirements
of the situation. The well-known CIA triad is the foundation security framework that can be used in any
9
environment, including the IoT environment. This framework helps to ensure responsible implementation
in the context of IoT security. The CIA triad framework is as follows.
1. Confidentiality. The access of information in the IoT environment needs to be limited to the user
and responsible agent.
2. Integrity. The data is not changed and manipulated. This involves ensuring that the data is accurate
and unchanged.
3. Availability. The information and system need to be available and not disrupted. This is to ensure
that authorized users have access to the data and information associated with it.
Figure 3. An illustration of threat model in IoT [2].
McNett et al. [7] suggested that holistic threat model should be constructed first, before implementing
any IoT solution. Threat modeling consists of six steps and is a great way to ensure that stakeholders
understand where their network's weaknesses lie. The threat modeling process follows these steps.
1. Asset identification
2. IoT system overview development.
3. Categorization of the IoT system.
4. Threat identification
5. Threat documentation
6. Threat rating
Xiao et al. [2] mentioned that IoT systems are vulnerable to physical, software, and network attacks
as well as privacy leakage. As shown in the Figure 3, some of the IoT security threats are as follows.
1. DoS attackers. Attackers send unwanted requests to targeted servers in order to prevent IoT devices
from obtaining services. One of the most dangerous kinds of DoS attacks is when a DDoS attacker
10
uses thousands of internet protocol addresses to request IoT services, and the servers have
legitimate access to IoT devices and attacks. Decentralized IoT devices with lightweight security
protocols are particularly vulnerable to DDoS attacks.
2. Jamming. The attacker sends fake signals to interrupt the ongoing radio transmissions of IoT
device, and during a failed communication attempt, the attacker can control the bandwidth, power,
central processing unit (CPU)of the IoT device or sensor.
3. Spoofing. A spoofing node uses an identity such as a MAC (Medium Access Control) address or
RFID tag to spoof his legitimate IoT device to gain unauthorized access to the IoT system and
launch further attacks such as DoS and man-in-the-middle attacks.
4. Man in the middle attack. The attackers send jamming and spoofing signals to stealthy monitor,
spy, and modify private communications between IoT devices.
5. Software attacks. Mobile malware such as worms and viruses can lead to financial losses, power
outages, and privacy leaks of IoT systems.
6. Privacy leakage. IoT systems should protect user privacy when caching and exchanging data.
Wearable devices that collect users' personal information, such as health and location data, create
a higher risk of privacy exposure.
B. Challenges with Previous Research
Conventional authentications are known for some critical challenges to solve. Due to its limitation in
hardware and firmware deployment, IoT may suffer to the complex environment, especially in a large scale
IoT network. Moreover, the key may leak in the security management procedures (for example in key
generation and distribution). Fang et al. specifies three significant challenges in supporting large scale IoT
applications that are as follows.
1. Long security induced latency. Traditional encryption methods require huge overhead and lengthy
processes to increase security levels, resulting in high communication and computational overhead
and, more importantly, high communication latency. As a result, large-scale IoT applications
require new authentication methods that are fast and lightweight.
2. Ineffective adaptation to dynamic IoT environment. Traditional security solutions rely on static
binary authentication process, which have difficulty learning and adapting to the dynamic IoT
environments they encounter, fail to continuously protect legitimate communications, and suffer
from poor security performance in addressing various security risks. Therefore, the new concepts
11
of continuous authentication and progressive authorization are very useful for optimizing the
overall system in a dynamic IoT environment.
3. Potential key leakage in the security procedures. Traditional cryptography requires proper key
management procedures to generate, distribute, update, and revoke security digital keys, which can
lead to key loss. This leads to huge attacker loopholes and pervasive security threats in large IoT
systems.
Debak et al [28] discuss that traditional IoT systems and computing platforms are not able to handle huge
data generated by the devices as well as harden the security layers. As a result, there is a significant loophole
in the security.
C. General AI Solution Process
The main task of device authentication is the classification task, to be more specific in classifying
legitimate users to access IoT devices. The AI solutions need to be able to accurately classify authorized
and unauthorized devices. Figure 4 shows the summary of the general AI solution process.
1. Data collection.
2. Data exploration and pre-processing.
3. Model selection.
4. Data conversion.
5. Training and testing.
6. Model evaluation and deployment.
Figure 4. General development process of AI or ML solutions
12
D. Recent Studies in AI based Authentication on IoT
Table 1 summarizes the studies based on the AI algorithms used in the IoT authentication system.
The table includes article or paper authors, year, title, review, and algorithms used. Some of the research
used the lightweight AI approach to overcome limited resource availability in an IoT environment. Certain
studies use various AI algorithms, and even combined with other highly researched domains such as
blockchain.
Table 1. Summary of recent research related to AI-based authentication on IoT environments based on the
algorithm used.
Literature
Reference
Review
Algorithm
[8], [17]
To overcome limited authentication and
authorization privileges for mobile users, this paper
proposes a lightweight method using smartcard
based to secure authentication process using neural
network model.
Neural Network
[10], [15]
To enhance the execution and encryption for
contractual agreements that solve security issues,
this paper discusses a blockchain based remote
mutual authentication that applies on smart devices
and cloud networks. Also, to overcome the risk of
leaking user privacy and data privacy risk, the paper
introduces a new technique, using a hashing
algorithm in signature verification scheme to
improve the security against insider attacks.
Blockchain based RMA
[14], [11]
To overcome IoT resource limitation and reliance
challenges, the paper proposes holistic
authentication and authorization techniques, where
machine learning based solution is adopted to
achieve adaptive access control. This is an
Support Vector Machine
(SVM)
13
interesting paper to discuss since it combines
theoretical and practical consideration.
[9]
To reduce human involvement as much as possible
as well as to standardize the development of
authentication agents without compromising the
robustness, security, and response time. This paper
presents an AI-enabled IoT based authentication
system.
Local Binary Pattern
Histogram based classifier
[39]
Some IoT authentication solutions are not practical
due to the small size of edge devices of IoT or not
cost efficient enough for mass adoption. This paper
proposes the use of Artificial Intelligence or
machine learning to address the issue of
authentication and authorization in edge devices
using fog computing model, with the application in
smart house problem. Considering multi-layer of AI
models and the complexity of the algorithm, this
might not be too practical in real IoT application
with limited resources.
Principal Component
Analysis (PCA), training
model is not specifically
discussed
[12]
IoT applications may cause serious challenges in
securing networks and data in transit. The paper
proposes a framework for a cloud based lightweight
cancelable biometric authentication system. This
paper is interesting to deep dive since it is quite
practical with current challenges by focusing on
biometric to authenticate.
CBS
14
IV. Discussion
This paper already discussed the common security challenges in IoT environment, which are device
authentication interception, DDoS attack, lack of intrusion detection, and lack of malware detection.
Agreeing to Fang et al., specific to authentication problems there are some main challenges due to current
technical limitation in resources. Therefore, lightweight AI based authentication is getting popular in recent
studies to overcome the security problems in IoT authentication.
The room for innovation in this domain is also huge. One of the interesting innovations is to combine
AI & blockchain technology to create a robust authentication system. Some of the reviewed literatures try
to build lightweight blockchain based system which is deployed on top of authentication system to
overcome the resource limitation on IoT.
Based on the literature review presented in this work, it can be concluded that most of the initial
research questions are answered. The summary of the analysis findings can be reviewed in Table 2.
Table 2. Summary of the analysis findings based on the research questions.
Research Question
Answer
What are the common challenges
in an IoT environment,
especially related to
authentication problems?
Due to its rapid development and resource limitation, IoT is known
for some security risks. Several security risks that are discussed in
this paper are device authentication interception, DDoS attack, lack
of intrusion detection, and lack of malware detection.
In this paper, specific to the authentication problems, there are four
main challenges with common IoT environment, which are as
follows.
1. Limited resources on IoT devices
2. Traditional encryption that impacts on latency
3. Difficulty in adapting to dynamic IoT environments
4. Potential key leakage in the procedures
15
How could AI help IoT devices
secure the authentication
process?
Recently AI has been known to be a solution for certain problems,
including to solve authentication problems. Lightweight AI is used to
create a biometric model for IoT devices to authenticate and authorize
users, as there is a limitation of resource in IoT device.
Reviewed papers discuss the use of face recognition, signature
verification, as well as smart card-based authentication. Along with
AI, improvement in encryption and hashing algorithms are also
discussed in the paper to protect confidentiality.
How is the recent research trend
in AI to solve the authentication
problems?
In authentication problems, the trend in practical study is to reduce
human involvement and effort as much as possible, as well as to
improve the authentication security without compromising the
response time, and robustness.
Another exciting research domain is the combination of AI and
blockchain to solve the authentication problem. The combination of
both is intended to provide robust level of encryption and execution
for contractual agreement that resolves interoperability and security
issues in IoT devices.
V. Conclusion
In this paper, literature on topics related to AI based authentication in IoT environments is studied.
One of the most common security challenges in IoT is authentication, where the risks of tampering,
counterfeiting, interception, and destruction in the process of information interaction and data transmission
between IoT nodes happen. Over the years, especially in the 2010s IoT technology is rapidly developing,
methods in improving IoT authentication security are also progressing significantly, and a lot of them are
involving AI as the solution.
From this literature review, it is understood that there are three main problems with the conventional
IoT authentication system, which are long security-induced latency, ineffective adaptation to dynamic IoT
environment, and potential key leakage in the security management procedures. From the reviewed studies
it is concluded that authentication systems that involve AI provide more robust security than conventional
16
authentication methods. This is reflected by the multi-layer security that is provided by the biometric-based
AI in the authentication services that prevent illegal access.
Moving forward, applicable research related to scalable authentication systems in IoT edges is
essential to solve IoT resource limitations. Moreover, the combination of AI and blockchain in
authentication system is an exciting topic to explore. This can be seen from the trend in recent papers, which
the combination of AI and blockchain provides robust level of encryption and execution that resolves
growing security issues.
17
References
1. H. Wu, H. Han, X. Wang and S. Sun, "Research on Artificial Intelligence Enhancing Internet of
Things Security: A Survey," in IEEE Access, vol. 8, pp. 153826-153848, 2020, doi:
10.1109/ACCESS.2020.3018170.
2. L. Xiao, X. Wan, X. Lu, Y. Zhang and D. Wu, "IoT Security Techniques Based on Machine
Learning: How Do IoT Devices Use AI to Enhance Security?," in IEEE Signal Processing
Magazine, vol. 35, no. 5, pp. 41-49, Sept. 2018, doi: 10.1109/MSP.2018.2825478.
3. G. Kornaros, "Hardware-Assisted Machine Learning in Resource-Constrained IoT Environments
for Security: Review and Future Prospective," in IEEE Access, vol. 10, pp. 58603-58622, 2022,
doi: 10.1109/ACCESS.2022.3179047.
4. Whitmore, Andrew, Anurag Agarwal, and Li Da Xu. “The Internet of Things—A Survey of
Topics and Trends.” Information systems frontiers 17.2 (2014): 261–274. Web.
5. Xiao. Y, Watson.W. (2019). “Guidance on Conducting a Systematic Literature Review”, Journal
of Planning Education and Research, Vol. 39(1) pp. 93–112, 2019.
6. F. Jindal, S. Mudgal, V. Choudhari and P. P. Churi, "Emerging trends in Internet of Things,"
2018 Fifth HCT Information Technology Trends (ITT), 2018, pp. 50-60, doi:
10.1109/CTIT.2018.8649535.
7. McNett, Jackie, Josh McNett, and Xiaoli Su. “IoT Security in Industry: A Threat Model of
Existing and Future Network Infrastructure.” Journal of applied security research ahead-of-
print.ahead-of-print (2022): 1–19. Web.
8. Alroobaea, R. et al. (2022) AI-assisted bio-inspired algorithm for secure IoT communication
networks. Cluster computing. [Online] 25 (3), 1805–1816.
9. Shelke, N. A. et al. (2022) ‘AI-enabled IoT based multimodal authentication system for securing
the hardware and software clients’, in 2022 International Conference on Machine Learning, Big
Data, Cloud and Parallel Computing (COM-IT-CON). [Online]. 2022 IEEE. pp. 545–550.
10. Chinnasamy, P. et al. (2022) Ciphertext-policy attribute-based encryption for cloud storage:
Toward data privacy and authentication in ai-enabled iot system. Mathematics (Basel). [Online]
10 (1), 68–.
11. Liang, Y. et al. (2020) Behavioral Biometrics for Continuous Authentication in the Internet-of-
Things Era: An Artificial Intelligence Perspective. IEEE internet of things journal. [Online] 7 (9),
9128–9143.
12. Punithavathi, P. et al. (2019) A lightweight machine learning-based authentication framework for
smart IoT devices. Information sciences. [Online] 484255–268.
18
13. Fang, H. et al. (2019) Fast Authentication and Progressive Authorization in Large-Scale IoT:
How to Leverage AI for Security Enhancement?
14. Mahbub, T. et al. (2020) ‘ANFIS based authentication performance evaluation for enhancing
security in Internet of Things’, in ACM International Conference Proceeding Series. [Online].
2020 ACM. pp. 1–4.
15. Fadlullah, Z. M. et al. (2022) Balancing QoS and Security in the Edge: Existing Practices,
Challenges, and 6G Opportunities with Machine Learning. IEEE Communications surveys and
tutorials. [Online] 1–1.
16. Mehri, V. A. et al. (2019) ‘Designing a Secure IoT System Architecture from a Virtual Premise
for a Collaborative AI Lab’, in [Online]. 2019 Blekinge Tekniska Högskola, Institutionen för
datalogi och datorsystemteknik. P.
17. Touqeer, H. et al. (2021) Smart home security: challenges, issues and solutions at different IoT
layers. The Journal of supercomputing. [Online] 77 (12), 14053–14089.
18. Zhang, S. et al. (2022) Blockchain and Federated Deep Reinforcement Learning Based Secure
Cloud-Edge-End Collaboration in Power IoT. IEEE wireless communications. [Online] 29 (2),
84–91.
19. Tsiatsis, V. et al. (2018) Internet of Things: Technologies and Applications for a New Age of
Intelligence. [Online]. San Diego, CA, USA: Elsevier Science.
20. Han, D. (2022) IoT Security in the Era of Artificial Intelligence. ProQuest Dissertations
Publishing.
21. Park, J. et al. (2020) ‘VibAuth: Enabling Accurate, Usable, and Deployable User Authentication
by Leveraging Vibration’, in 2020 International Conference on Information and Communication
Technology Convergence (ICTC). [Online]. 2020 IEEE. pp. 1771–1773.
22. Sudharsan, B. et al. (2022) Smart speaker design and implementation with biometric
authentication and advanced voice interaction capability.
23. Huyghue, B. D. (2021) Cybersecurity, Internet of Things, and Risk Management for Businesses.
ProQuest Dissertations Publishing.
24. Priyanka Bothra et al. (2021) How Can Applications of Blockchain and Artificial Intelligence
Improve Performance of Internet of Things? -- A Survey. arXiv.org.
25. Fang, He, Angie Qi, and Xianbin Wang. “Fast Authentication and Progressive Authorization in
Large-Scale IoT: How to Leverage AI for Security Enhancement.” IEEE network 34.3 (2020):
24–29. Web.
26. Deebak, B. D. et al. (2022) Lightweight Blockchain Based Remote Mutual Authentication for AI-
Empowered IoT Sustainable Computing Systems. IEEE internet of things journal. [Online] 1–1.
19
27. Punithavathi, P. et al. (2019) A lightweight machine learning-based authentication framework for
smart IoT devices. Information sciences. [Online] 484255–268.
28. Deebak, B. D. & AL-Turjman, F. (2021) Lightweight authentication for IoT/Cloud-based
forensics in intelligent data computing. Future generation computer systems. [Online] 116406–
425.
29. N. Koroniotis, N. Moustafa, F. Schiliro, P. Gauravaram and H. Janicke, "A Holistic Review of
Cybersecurity and Reliability Perspectives in Smart Airports," in IEEE Access, vol. 8, pp.
209802-209834, 2020, doi: 10.1109/ACCESS.2020.3036728.
30. R. Milton, D. Hay, S. Gray, B. Buyuklieva and A. Hudson-Smith, "Smart IoT and Soft AI,"
Living in the Internet of Things: Cybersecurity of the IoT - 2018, 2018, pp. 1-6, doi:
10.1049/cp.2018.0016.
31. D. Man, F. Zeng, J. Lv, S. Xuan, W. Yang and M. Guizani, "AI-based Intrusion Detection for
Intelligence Internet of Vehicles," in IEEE Consumer Electronics Magazine, doi:
10.1109/MCE.2021.3137790.
32. A. Bandi, "A Review Towards AI Empowered 6G Communication Requirements, Applications,
and Technologies in Mobile Edge Computing," 2022 6th International Conference on Computing
Methodologies and Communication (ICCMC), 2022, pp. 12-17, doi:
10.1109/ICCMC53470.2022.9754049.
33. C. Dou et al., "Challenges of emerging memory and memristor based circuits: Nonvolatile logics,
IoT security, deep learning and neuromorphic computing," 2017 IEEE 12th International
Conference on ASIC (ASICON), 2017, pp. 1survey40-143, doi:
10.1109/ASICON.2017.8252431.
34. V. Hassija, V. Chamola, V. Gupta, S. Jain and N. Guizani, "A Survey on Supply Chain Security:
Application Areas, Security Threats, and Solution Architectures," in IEEE Internet of Things
Journal, vol. 8, no. 8, pp. 6222-6246, 15 April15, 2021, doi: 10.1109/JIOT.2020.3025775.
35. C. Duru, J. Ladeji–Osias, K. Wandji, T. Otily and R. Kone, "A Review of Human Immune
Inspired Algorithms for Intrusion Detection Systems," 2022 IEEE World AI IoT Congress
(AIIoT), 2022, pp. 364-371, doi: 10.1109/AIIoT54504.2022.9817213.
36. Srinivasa Rao Gundu; Charanarur Panem; S. Satheesh, "High‐Performance Computing‐Based
Scalable “Cloud Forensics‐as‐a‐Service” Readiness Framework Factors—A Review," in Cyber
Security and Network Security , Wiley, 2022, pp.27-45, doi: 10.1002/9781119812555.ch2.
37. R. O. Andrade, S. G. Yoo, L. Tello-Oquendo and I. Ortiz-Garcés, "A Comprehensive Study of
the IoT Cybersecurity in Smart Cities," in IEEE Access, vol. 8, pp. 228922-228941, 2020, doi:
10.1109/ACCESS.2020.3046442.
20
38. R. Bertoglio, C. Corbo, F. M. Renga and M. Matteucci, "The Digital Agricultural Revolution: A
Bibliometric Analysis Literature Review," in IEEE Access, vol. 9, pp. 134762-134782, 2021, doi:
10.1109/ACCESS.2021.3115258.
39. Zareen, M. S. et al. (2019) ‘Artificial Intelligence/ Machine Learning in IoT for Authentication
and Authorization of Edge Devices’, in 2019 International Conference on Applied and
Engineering Mathematics (ICAEM). [Online]. 2019 IEEE. pp. 220–224.
40. B. Ikharo, A. Obiagwu, C. Obasi, S. U. Hussein and P. Akah, "Security for Internet-of-Things
Enabled E-Health using Blockchain and Artificial Intelligence: A Novel Integration Framework,"
2021 1st International Conference on Multidisciplinary Engineering and Applied Science
(ICMEAS), 2021, pp. 1-4, doi: 10.1109/ICMEAS52683.2021.9692368.
41. H. K. Bharadwaj et al., "A Review on the Role of Machine Learning in Enabling IoT Based
Healthcare Applications," in IEEE Access, vol. 9, pp. 38859-38890, 2021, doi:
10.1109/ACCESS.2021.3059858.
42. B. A. Salau, A. Rawal and D. B. Rawat, "Recent Advances in Artificial Intelligence for Wireless
Internet of Things and Cyber–Physical Systems: A Comprehensive Survey," in IEEE Internet of
Things Journal, vol. 9, no. 15, pp. 12916-12930, 1 Aug.1, 2022, doi:
10.1109/JIOT.2022.3170449.
43. A. Chakraborty, R. Bagavathi and U. Tomer, "A Comprehensive Decomposition towards the
Facets of Quality in IoT," 2020 International Conference on Smart Electronics and
Communication (ICOSEC), 2020, pp. 759-764, doi: 10.1109/ICOSEC49089.2020.9215428.
44. Petersen, Kai, Sairam Vakkalanka, and Ludwik Kuzniarz. “Guidelines for Conducting
Systematic Mapping Studies in Software Engineering: An Update.” Information and software
technology 64 (2015): 1–18. Web.
45. Xiao. Y, Watson.W. (2019). “Guidance on Conducting a Systematic Literature Review”, Journal
of Planning Education and Research, Vol. 39(1) pp. 93–112, 2019.
